Getting Unstuck
AI and the UK in the 2020s
Getting Unstuck: AI and the UK in the 2020s
The following essay was written for the TxP Progress Prize, an essay competition on the theme of tech policy solutions. The prompt begins “Britain is stuck. A new movement around science, technology and economic progress offers a way forward…”.
In 2018, Ian Hogarth, now head of the UK AI Task Force, predicted that “AI policy will become the single most important area of government policy.” There are numerous policy areas that will determine the UK’s trajectory in the coming decades, from the sci-fi of biotech to the minutiae of planning permission. However, as a general technology, AI will enable progress in all of these fields, and is therefore the focus of this essay.
Much is made of multi-billion dollar AI unicorns, like OpenAI and Anthropic, and it could be tempting to argue that UK AI policy should prioritise the enabling of AI businesses. However, we might refer back to the cautionary tale of Quaero, an initiative that absorbed hundreds of millions of government funds attempting to build a “European Google”. Or the fact that Silicon Valley has done little to improve the lot of most Americans on metrics from the HDI to life expectancy. We argue that the UK government would be better served focussing on access and utilisation of AI across the population — rather than chasing a golden goose, UK policy should prioritise getting AI solutions into people’s hands, realising tangible benefits.
The prompt states that Britain is stuck. If so, we suggest two unsticking priorities: getting going, and staying on track.
Let’s Get Moving
To address the first question, of economic growth and better utilisation, we identify three key groups: large companies/multinationals, small and medium enterprises (SMEs) and the public sector. Given the ample resources and legions of consultants aiding AI transformation at large companies, we’ll focus on empowering SMEs and the public sector.
Data Management Taskforce
The UK is a nation of shopkeepers; Small Medium Enterprises (SMEs) employ 2/3rds of British workers and account for half of all business income. However, most SMEs lack the basic foundations of good data management to enabled AI solutions. We propose a Data Management Taskforce to advise and support SMEs in organising data to be able to make use of AI tools. The remit could include using AI to manage emails and customer support, moving documents into cloud repositories to enable Retrieval Augmented Generation (RAG), or setting up a receipt system that make use of AI accounting.
Skilling Up
The UK is in the midst of a decade-long skills and productivity plateau. Since 2003, real investment in adult skills and lifelong learning has fallen over 30%. To maximise the benefits of AI, employees will need training to adapt to new working paradigms, as well as re-skill as tasks are automated. The government should launch a learning platform similar to Denmark’s AMU, alongside a rebate for private investment in skills. The UK has a legacy of painful labour shifts dating back to the loom-smashing of the Luddites; in the 2020s we should aim for smooth, equitable transitions, keeping people productive and in work.
AI For Government
A US executive order on AI has driven a hiring spree of 400 AI Officers to spearhead AI transformation in federal agencies. The UK should be equally ambitious, recruiting AI expertise into government departments. This will strengthen the capacity of the government to innovate with deep-tech, a forgotten capacity of the Entrepreneurial State, and decrease dependency on private-sector firms and consultancies. In doing so, the Civil Service should refresh its pay structure, increasing (and fully using) the range of pay bands to attract and retain top talent.
UK Data Exchange
A recent lawsuit by the New York Times against OpenAI reflects growing concern about the role of creatives facing corporations that mine the internet for intellectual property to train AI models. We propose a nationalised self-sovereign data exchange — operated by an independent public-private body — where creative professionals or organisations can submit content and receive royalties every time their content is used in AI training. Many companies are anxious about the legal uncertainty of frontier AI models (hence the introduction of copyright shields). The certainty of using a platform of this kind would be preferable to the current legal grey area occupied by AI systems, as well as offering some compensation for the people whose work powers AI.
Staying on Track
Second, we address broader societal impacts of AI, and protections.
AI and Data Literacy
Beyond equipping workers with the relevant skills to thrive in a data and AI driven economy, there is a pressing need to inoculate the broader population, especially young people, with the know-how to navigate an information-scape that is saturated with misinformation and increasingly exposed to AI generated content. The UK should follow the footsteps of countries like Finland by rolling out an AI and Data Literacy program at schools, in order to better equip children with skills to scrutinise the veracity and quality of data they encounter in their daily lives.
Protections for Workers
The impact on employment as a result of widespread AI adoption is potentially massive - this includes the disruption to labour markets, but also on the ways that people work. How AI is brought into the workplace needs to be carefully considered, with protections for workers rights in the foreground. For example, in light of regulatory ambiguity on AI in the UK currently, we propose that employment legislation is updated to provide safeguards against discriminatory decision making made by AI models and that there is employee right of refusal for the use of high risk AI models in the workplace.
Content Provenance
Much has been made of the risk of generative AI to exacerbate breakdowns in cross-societal trust, as we lose the ability to tell what is real or generated. These issues could be partially addressed by content provenance and watermarking systems, such as C2PA, which enables anyone to view the source of, and subsequent edits to, an image. The government should gradually introduce requirements for content provenance from news outlets and broadcasters, as well as requiring AI tools and software sold in the UK to include provenance features.
In conclusion, we advocate a focus on tangible benefits — AI as a tide lifting all ships. AI is not an end unto itself, but a powerful collection of tools which can and should underpin major revitalisation for the UK. We envision a future where the UK not only competes but excels in the global landscape, powered by shared public value. The time is ripe for the UK to embark on a moonshot approach, securing a pivotal role in an AI-driven future.
Sam Meeson is a Researcher at Evident. He is currently studying for a PhD in Digital Anthropology at UCL.
Alex Inch is a Data Scientist at Evident. He previously worked at Tortoise Media on the Global AI Index.
 | A guest post by
|